Gregory Eady

Project title

The Discourse Project: Investigating and Improving the Quality of Political Discourse

What is your project about?

Politicians and citizens regularly criticize the political discourse as unconstructive, uncivil, and partisan. How can we have more constructive political conversations? This project will seek to answer that question through a series of experiments using new AI tools that can take part in conversations. These tools will help test which types of approaches to argumentation lead to more constructive, more civil, and less polarized discussions. The project will also study how people's identities - like their political background, gender, or race - can affect the quality of conversations. Finally, it will test whether using high-quality evidence and information about policy from official or government sources can help reduce potentially harmful beliefs and build trust in democracy.

How did you become interested in your particular field of research?

As someone who conducts research on social media, I've long found it frustrating that so much political discussion online is unproductive, polarizing, or hostile. But studying why this happens - and how to fix it - is methodologically difficult. To test what makes a conversation more constructive, we need to be able to vary how people talk and then see what changes as a result. That kind of controlled comparison is hard to do in real conversations. But recent advances in AI make it possible to simulate dialogue and test how different approaches to argumentation affect the quality of discussion. This opens up entirely new possibilities for studying and improving how people talk politics online.

What are the scientific challenges and perspectives in your project?

This project faces both theoretical and technical challenges. A core scientific task is to identify which specific features of dialogue - reasoning style, use of evidence, tone, engagement with counter-arguments - most reliably lead to more constructive political discussion. It is equally important to understand how people can be encouraged to adopt these approaches in practice. Technically, the project must also develop ways to equip AI systems to reference and argue using official and governmental sources, in order to test whether use of the highest quality evidence in dialogue can shift political attitudes and increase democratic trust. Addressing these challenges requires careful experimental design, new applications of AI, and close attention to how people respond to different forms of political argument.

What is your estimate of the impact, which your project may have to society in the long term?

The goal of this project is to uncover approaches to political dialogue that people can use if they want to have more informed, respectful, and constructive conversations about politics. In the long term, the findings will help individuals, educators, journalists, and technology platforms design better ways to promote healthier political discussion.

Which impact do you expect the Sapere Aude programme will have on your career as a researcher?

The Sapere Aude programme will allow me to build a research group and run large-scale experiments on political discourse that would otherwise be out of reach. It will strengthen my work at the intersection of political behavior and computational methods, and deepen my collaborations with both international and Danish researchers. Finally, the project will expand my methodological expertise in using AI for experimental research and provide a strong foundation for long-term contributions to the study of democratic resilience.

Background and personal life

I grew up in the suburbs of Ontario, Canada, and studied political science at the University of Toronto. After completing my PhD, I was a post-doctoral research at New York University, where I researched the effects of social media on political attitudes and behavior. In 2019, I moved to Denmark to continue this line of research, contribute to building the Social Data Science program at the University of Copenhagen, and teach quantitative methods to master's students in political science.